ipynb to py conversion process added

mnist.ipynb

@@ -145,6 +145,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
+    "# model definition\n",
     "def linear_classifier():\n",
     "    return nn.Sequential(\n",
     "        Reshape((-1, 784)),\n",
@@ -209,7 +210,11 @@
   {
    "cell_type": "code",
    "execution_count": 31,
-   "metadata": {},
+   "metadata": {
+    "tags": [
+     "exclude"
+    ]
+   },
    "outputs": [],
    "source": [
     "# with open('./linear_classifier.pkl', 'wb') as model_file:\n",
@@ -218,33 +223,47 @@
   },
   {
    "cell_type": "markdown",
-   "metadata": {},
+   "metadata": {
+    "tags": [
+     "exclude"
+    ]
+   },
    "source": [
     "#### commit to .py file for deployment"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 33,
-   "metadata": {},
+   "execution_count": 43,
+   "metadata": {
+    "tags": [
+     "exclude"
+    ]
+   },
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
       "[NbConvertApp] Converting notebook mnist.ipynb to script\n",
-      "[NbConvertApp] Writing 6388 bytes to mnist.py\n"
+      "[NbConvertApp] Writing 3123 bytes to mnist.py\n"
      ]
     }
    ],
    "source": [
-    "!jupyter nbconvert --to script mnist.ipynb"
+    "# !jupyter nbconvert --to script mnist.ipynb\n",
+    "!jupyter nbconvert --to script --TagRemovePreprocessor.remove_cell_tags=\"exclude\" --TemplateExporter.exclude_input_prompt=True mnist.ipynb\n",
+    "\n"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": null,
-   "metadata": {},
+   "metadata": {
+    "tags": [
+     "exclude"
+    ]
+   },
    "outputs": [],
    "source": [
     "# from IPython.display import HTML, display, Image\n",

mnist.py

@@ -1,9 +1,6 @@
 #!/usr/bin/env python
 # coding: utf-8
 
-# In[1]:
-
-
 import torch
 from torch import nn
 import torch.nn.functional as F
@@ -19,17 +16,11 @@ get_ipython().run_line_magic('matplotlib', 'inline')
 plt.rcParams['figure.figsize'] = [2, 2]
 
 
-# In[2]:
-
-
 dataset_nm = 'mnist'
 x,y = 'image', 'label'
 ds = load_dataset(dataset_nm)
 
 
-# In[3]:
-
-
 def transform_ds(b):
     b[x] = [TF.to_tensor(ele) for ele in b[x]]
     return b
@@ -38,9 +29,6 @@ dst = ds.with_transform(transform_ds)
 plt.imshow(dst['train'][0]['image'].permute(1,2,0));
 
 
-# In[4]:
-
-
 bs = 1024
 class DataLoaders:
     def __init__(self, train_ds, valid_ds, bs, collate_fn, **kwargs):
@@ -56,9 +44,6 @@ xb,yb = next(iter(dls.train))
 xb.shape, yb.shape
 
 
-# In[5]:
-
-
 class Reshape(nn.Module):
     def __init__(self, dim):
         super().__init__()
@@ -68,9 +53,6 @@ class Reshape(nn.Module):
         return x.reshape(self.dim)
 
 
-# In[6]:
-
-
 def cnn_classifier():
     ks,stride = 3,2
     return nn.Sequential(
@@ -89,9 +71,7 @@ def cnn_classifier():
     )
 
 
-# In[7]:
-
-
+# model definition
 def linear_classifier():
     return nn.Sequential(
         Reshape((-1, 784)),
@@ -103,9 +83,6 @@ def linear_classifier():
     )
 
 
-# In[8]:
-
-
 model = linear_classifier()
 lr = 0.1
 max_lr = 0.1
@@ -133,128 +110,7 @@ for epoch in range(epochs):
     
 
 
-# In[31]:
-
-
-# with open('./linear_classifier.pkl', 'wb') as model_file:
-#     pickle.dump(model, model_file)
-
-
-# #### commit to .py file for deployment
-
-# In[32]:
-
-
-get_ipython().system('jupyter nbconvert --to script mnist.ipynb')
-
-
-# In[29]:
-
-
-get_ipython().system('pip3 install markupsafe')
-
-
-
-# In[ ]:
-
-
-# from IPython.display import HTML, display, Image
-# %%html
-# <style>
-#     #whiteboard {
-#         border: 3px solid black;
-#         border-radius: 6px;  
-#         background-color: #FFFFFF;
-#     }
-#     #capture-button {
-#         background-color: #3F52D9; 
-#         color: white;
-#         border: none;
-#         padding: 10px 20px;
-#         cursor: pointer;
-#         font-size: 16px;
-#         border-radius: 3px;
-#         margin-top: 10px;
-#         width: 190px;
-#         margin-right: 20px;
-#     }
-#     #clear-button {
-#         background-color: #FF0000,; 
-#         color: black;
-#         border: none;
-#         padding: 10px 20px;
-#         cursor: pointer;
-#         font-size: 16px;
-#         border-radius: 3px;
-#         margin-top: 10px;
-#         width: 190px;
-#     }
-#     #container {
-#       display: flex;
-#       flex-direction: column; /* Arrange children vertically */
-#       align-items: center; /* Center horizontally */
-#       justify-content: center;
-#     }
-#     #btn-container {
-#       display: flex;
-#       flex-direction: row; /* Arrange children vertically */
-#       align-items: center; /* Center horizontally */
-#     }
-
-# </style>
-# <div id='container'>
-# <canvas id="whiteboard" width="400" height="200" fill_rect='white'></canvas>
-# <div id='btn-container'>
-# <button id="capture-button">Predict</button>
-# <button id="clear-button">Clear</button>
-# </div>
-
-# </div>
-# <script>
-#     var canvas = document.getElementById('whiteboard');
-#     var context = canvas.getContext('2d');
-#     var drawing = false;
-#     canvas.addEventListener('mousedown', function (e) {
-#         drawing = true;
-#         context.beginPath();
-#         context.moveTo(e.clientX - canvas.getBoundingClientRect().left, e.clientY - canvas.getBoundingClientRect().top);
-#     });
-#     canvas.addEventListener('mousemove', function (e) {
-#         if (drawing) {
-#             context.lineTo(e.clientX - canvas.getBoundingClientRect().left, e.clientY - canvas.getBoundingClientRect().top);
-#             context.stroke();
-#         }
-#     });
-#     canvas.addEventListener('mouseup', function () {
-#         drawing = false;
-#     });
-#     canvas.addEventListener('mouseout', function () {
-#         drawing = false;
-#     });
-    
-#     var clearButton = document.getElementById('clear-button');
-#     clearButton.addEventListener('click', function () {
-#        context.clearRect(0, 0, canvas.width, canvas.height);
-#     });
-
-#    var captureButton = document.getElementById('capture-button');
-#     captureButton.addEventListener('click', function () {
-#        // Convert the canvas content to a data URL (image)
-#     var imageData = canvas.toDataURL("image/png");
-
-#     // Send the image data to the Jupyter kernel variable
-#     IPython.notebook.kernel.execute('image_data = "' + imageData + '"');
-#     });
-# </script>
-
-
-# In[ ]:
-
-
-
-
 
-# In[ ]: